The conversation view

The conversation view is where you talk to Mel's agent: it replaces the pane's terminal while a conversation is open, and shows your prompts, the answers, every tool call, and the cards that ask for your approval or a decision. It keeps the terminal aesthetic — everything renders in monospace — but the content is a structured transcript, not raw scrollback.

The landing page

A fresh conversation opens bottom-anchored: the heading "Start a conversation with Mel", the line "Describe a task below and Mel will work in <your directory>", and a Recent activity box listing your latest conversations, which you can click to resume (View all opens /conversations; before your first conversation the box offers starter prompts instead). Below the input, a hint row reminds you of the essentials — / for commands, shift+↵ for a new line, Ctrl+Shift+Y (⌃⇧Y on a Mac) to open a conversation.

Writing the prompt

The input is a real multi-line editor, not a single line.

• ↵ sends. shift+↵ starts a new line — as do alt+↵ and ctrl+J, so whichever habit you arrive with works.
• Long prompts wrap; they never scroll sideways.
• The box grows as you write, up to about ten lines, then scrolls inside itself. It never grows so far that the transcript above it disappears.
• Pasting a multi-line block keeps its line breaks.
• ctrl+G (or the ⤢ button at the right of the input) pops the prompt out into a larger compose box for something long. esc closes it and keeps what you wrote.
• ↑ opens the prompt/command history — but only when the caret is on the first line. Anywhere else it moves the caret up, like any other editor.
• Your unsent draft stays put when you switch panes, and comes back if you leave a conversation and reopen it.
• Sending while the agent works. A message you send while Mel is thinking joins the conversation right away. Shell commands and slash commands wait for the current step to finish.
• Mermaid diagrams in an answer or a plan are drawn as pictures, on your computer: enlarge one, see its code, or save it as PNG or PDF.

What appears in the transcript

• User turns — your prompt on a line with a status dot, in monospace.
• Streaming response — the assistant's answer streams in as markdown, rendered in monospace with code blocks and formatting intact.
• Thoughts — models that expose reasoning stream it into a collapsible "Thoughts" section. It's open while the model is thinking and collapses to a "Thought for N s" line when done; click to expand it again.
• Tool cards — each tool call is a full-width row: the tool name and target (a file path, a search pattern, a command), a running indicator while it executes, a checkmark when done, and expandable output. See Tools & approvals.
• Tasks card — the agent's live checklist for multi-step work. See Tasks & the agent lifecycle.
• Diff cards — every applied file change, with +/- counts and red/green rows.
• Approval cards — inline gates for writes and commands.
• Question cards — when the agent needs a decision only you can make, it asks with numbered options and a recommended choice.
• Pictures, audio and video the agent makes appear in the conversation.

Agent questions

When the agent needs a decision only you can make, it stops and shows an Agent questions card instead of guessing. The run waits, and the conversation is marked as needing you (green, like an approval).

• The card shows the question and one numbered row per option. The option the agent would pick carries a grey Recommended badge.
• Click a row to select it, then click Next (or press Enter while the message box is empty) to send it.
• Type your own answer in the last row if none of the options fits.
• Skip all lets the agent decide; it is told "Skipped — use your best judgement."
• Once answered, the card's title changes to You answered and keeps your choice in the transcript. If the turn ends before you answer (you stopped it or sent a new prompt), it says Question left unanswered.

A question can be asked even in autonomous mode: permission to act is not the same as knowing what you want.

Running shell commands mid-conversation

You don't have to leave the conversation to use the shell:

• Prefix your input with ! to force it to run as a shell command.
• Or just type the command — Mel's classifier notices shell-shaped input, shows an "autodetected shell command" banner, and runs it in the shell. Press Ctrl+I (⌃I on a Mac) to override the detection either way, or turn detection off in Settings → AI → Autodetect terminal commands. See Shell or agent.

The command runs in the pane's working directory and the result persists in the transcript as a shell block (❯ command plus output and exit status). A successful cd even updates the pane's tracked directory. Typing claude, codex, ssh, or vim here launches a full interactive session embedded in the conversation — see Embedded CLI agents.

After a turn: feedback and usage

Each completed response ends with a feedback row (👍 👎, Copy response) and a usage line — the model, tokens, what the answer cost in credits (0 credits · your key on your own key) and time. When Mel does not know the cost, it shows none rather than a made-up zero. The conversation's running total is shown too. Expanding the line shows:

• Usage summary — the model used, why the router chose it, tokens in/out, and context window used (with the circular gauge).
• Tool call summary — files changed, diffs applied, commands executed.
• Last response time — time to first token and total turn time.

The same circular context gauge lives in the input footer; its percentage is printed beside it from 2% on, and hovering it explains what it measures. When the window gets tight, Mel compacts the conversation — see Conversation management.

FAQ

Why did the agent stop and ask me a question? It reached a choice it can't make for you, such as which of two designs you want. Pick an option, type your own answer, or press Skip all to let it decide.

Can I send a message while the agent is working? Yes. Enter sends it into the running turn; Ctrl+Enter (⌘↵ on a Mac) queues it for when the turn finishes.

Where do I see what a turn cost? On the usage line under the answer. Expand it for the model, tokens and timing.